SRVLOGIC-1151 | Add run-with-retry helper for resilient Go dependency resolution - #457
jankrystof-ibm wants to merge 1 commit into
Conversation
|
After the CI run I see
|
|
Adding @fantonangeli to reviewers, who is better with Javascript. |
fantonangeli
left a comment
There was a problem hiding this comment.
LGTM, I would just double check the lockfile
| jest: | ||
| specifier: ^29.7.0 | ||
| version: 29.7.0(@types/node@24.13.3)(node-notifier@8.0.2)(ts-node@10.9.2(@swc/core@1.3.92)(@types/node@24.13.3)(typescript@5.9.3)) | ||
| version: 29.7.0(@types/node@26.1.1)(node-notifier@8.0.2)(ts-node@10.9.2(@swc/core@1.3.92)(@types/node@26.1.1)(typescript@5.9.3)) |
There was a problem hiding this comment.
Are all these @types/node 24.13.3 → 26.1.1 lockfile changes should not be expected.
Can you try to re-generate it?
There was a problem hiding this comment.
Yes, those types need to stay as before, aligned with the supported node version.
There was a problem hiding this comment.
@jankrystof-ibm could you please take a look?
There was a problem hiding this comment.
I think we may leave it as it is. I have the same problem when regenerating the lock file. Those types/node should be compatible. We have ^24.0.11 in the package.json and these are supposed to be compatible with it.
ticket: https://redhat.atlassian.net/browse/SRVLOGIC-1151
CI bootstrap occasionally aborts when Go dependency resolution (go work sync / go mod tidy) hits a transient network failure, e.g. checksum verification against sum.golang.org. The newly introduced package wraps the resolution step in the install hooks with a bounded retry+backoff wrapper so a transient outage no longer kills the whole job immediately but it enables several retries.
New package
@kie-tools-scripts/run-with-retryensures 5 attempts with backoff 5/15/30/60s between attempts (last value repeats, so raising--retrieskeeps spacing later attempts by a minute). On exhaustion it exits with the original exit code, so non-transient failures still surface and are not masked.